Spherical Hashing in TensorFlow

Spherical Hashing [1] is a data-dependent hashing function which learns binary codes for data points by partitioning the feature space with hyperspheres, rather than hyperplanes.

Requirements

• Python 3.5+
• TensorFlow 1.6+

References

1. Heo, Jae-Pil, et al. "Spherical hashing." 2012 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2012.